Beyond Raw Backscatter: Multiscale Feature Extraction from Elastic Lidar Observations

Elastic backscatter lidar and ceilometer systems provide continuous observations of aerosol and cloud vertical structure, but the interpretation of conventional attenuated backscatter products is often limited by the dominance of signal amplitude, strong event-to-event variability, and the reduced visibility of subtle internal features. In this study, we present a refinement framework designed to extract additional structural information from elastic lidar measurements through multiscale local diagnostics applied directly to the native backscatter field. The methodology combines standardized residual fields, local gradients, variance-based metrics, space–time decorrelation scales and structure functions to highlight atmospheric boundaries, internal layering, mixing zones, and coherent structures that are not always evident in conventional representations. The approach is evaluated through three contrasting atmospheric case studies observed in 2024. Two spring events are associated with mineral dust intrusions characterized by different vertical coupling with the planetary boundary layer, while a summer case represents a non-dust regime dominated by diurnal boundary-layer evolution. The refined diagnostics consistently reveal features hidden or only weakly visible in the raw backscatter field, including sharp interfaces, embedded stratification, wave-like perturbations and transitions between decoupled and mixed atmospheric states. Results show that the proposed metrics enable a more objective description of aerosol-layer dynamics and boundary–layer interactions without requiring complex inversion procedures or auxiliary measurements. Because the method relies only on standard elastic lidar observations, it is in principle applicable to ceilometer and lidar monitoring networks. However, the present evaluation is based on three contrasting case studies and should therefore be regarded as a proof-of-concept demonstration. The framework offers a candidate pathway for enhanced atmospheric feature detection and improved interpretation of routine profiling observations, with automated regime classification as a longer-term goal requiring validation on larger and more diverse datasets.